Server Farms Research Articles

Web Services Clustering via Exploring Unified Content and Structural Semantic Representation

Clustering Web services can improve the quality and efficiency of service discovery and management within a service repository. Nowadays, Web services frequently interact (e.g., composition relation and tag sharing relation) with each other to form a complex and heterogeneous service relationship network. The rich network relations inherently reflect either positive or negative clustering association between Web services, which can be a strong supplement to service semantics for characterizing functional affinities between Web services. In this paper, we propose to cluster Web services by utilizing both description documents and the structural information from the service relationship network. We first learn the content semantic information from service description documents based on the widely used Doc2vec model, and meanwhile, learn the structural semantic information from the service relationship network based on a network representation learning algorithm. Then, we propose to pretrain the content and structural semantic information to obtain the most relevant and unified features through training a service classification model with partially labeled data. Finally, a spectral clustering algorithm is utilized for Web services clustering based on the above unified features with preserved content and structural semantics. Therefore, the proposed services clustering approach takes advantage of both service content semantic and service network structure semantic based similarity between services. Extensive experiments are conducted on a real-world dataset from ProgrammableWeb, composed of 12919 Web API services. Experimental results demonstrate that our approach yields an improvement of 4.78% in precision and 5.4% in recall over the state-of-the-art method.

A Non-Markovian Queue with Second Optional Service, Disasters, Working Breakdowns and Working Vacation

Inside this research article, we look at a non-Markovian queue (M/G/1) with second optional service, disaster events, working breakdowns and working vacation. There are two servers in the system: a primary and a backup. First Essential Service (FES) is delivered by a primary (backup) server to all arriving customers. A customer may choose the Second Optional Service (SOS) after his initial service is finished. When a disaster happens, all customers are made to evacuate the system, as well as the primary server crashes. The primary server is dispatched to repair at the first sign of a breakdown and the repair period begins right away. While a primary server is being repaired, a backup server is servicing customers at such a reduced rate. If a system is inactive while it is operating, the primary server will just go on vacation. The primary server, which is on working vacation mode and serving at a reduced rate, subsequently serves the new customers. The disasters have no effect on working vacation period. The supplementary variable methodology is used to determine the probability generating function (PGF) of the number of customers throughout typical peak times, working breakdown periods and working vacation periods, as well as specific metrics of effectiveness. Some statistical outcomes are shown at the end.

Develop a New Lightweight Consensus Algorithm for IoT

Blockchain has gained a lot of attention in non-financial applications like the Internet of Things (IoT). It is characterized by security, decentralization, anonymity, and auditability. Today's technology makes extensive use of the internet of things to gather data from sensors and carry out specific tasks by processing it according to the requirements. To produce secure and impermeable records of transactions, the collected data can be transferred through the blockchain network. The integration of blockchain with IoT has enormous promise due to its ability to offer decentralized computing, storage, and exchange for IoT data. Conventional blockchains generally have low throughput, long transaction delays, and rely on computationally expensive consensus algorithms. However, because of its limitations, IoT applications need a blockchain network that has a lightweight consensus algorithm to integrate with it successfully. In this paper, we implement and design a Lightweight Practical Byzantine Fault Tolerance (LPBFT) consensus algorithm that improves the conventional (PBFT) algorithm. Through the implementation of the proposed LPBFT, it has shown: a simpler message structure; doesn't need a primary server; low computation time; and consists of just two stages (request and reply). That makes it workable efficiently in IoT networks.

Hybrid Edge Cloud: A Pragmatic Approach for Decentralized Cloud Computing

The massive growth of connected devices, including sensors and machines, is revolutionizing every aspect of human life. The socioeconomic impacts are significant and have already transformed many industries. In this article, we discuss some of the challenges of the explosion of devices and exponential growth in personal digital content and machine generated data. The centralized cloud architecture adopted in the early days of mobile Internet was designed primarily to allow access to data stored on the worldwide web. Today, use cases have evolved significantly. Humans and devices are now producing most of the data consumed on the Internet. As a result, the existing centralized cloud infrastructure is no longer efficient or sustainable. There is significant waste of network bandwidth to send terabytes of data to server farms that may be hundreds of kilometers away from the source and/or destination of data. The prevalent centralized cloud architecture does not adequately leverage massive amounts of computing resources on smart devices, which are idle most of the time. Moreover, despite all the efforts to reduce network latencies, communication to server farms is a major bottleneck for latency in many applications. We introduce a novel architectural approach to cloud decentralization called hybrid edge cloud (HEC) that minimizes network bandwidth usage, reduces communication latencies, and leverages resources on smart devices to reduce the burden on server farms and other centralized computing resources. HEC combines the benefits of new network technologies such as 5G and WiFi 6 in private and public clouds to leverage computing resources on smart devices to build a sustainable decentralized infrastructure for the hyper-connected world.

The Cruel Optimism of Anthropocene Technologies: Suspicion and Fascination of Technology in Okja, What Happened to Monday, and Geostorm

The Cruel Optimism of Anthropocene Technologies: Suspicion and Fascination of Technology in <i>Okja</i>, <i>What Happened to Monday</i>, and <i>Geostorm</i>

Parallelization of Array Method with Hybrid Programming: OpenMP and MPI

For parallelization of applications with high processing times and large amounts of storage in High Performance Computing (HPC) systems, shared memory programming and distributed memory programming have been used; a parallel application is represented by Parallel Task Graphs (PTGs) using Directed Acyclic Graphs (DAGs). For the execution of PTGs in HPC systems, a scheduler is executed in two phases: scheduling and allocation; the execution of the scheduler is considered an NP-complete combinatorial problem and requires large amounts of storage and long processing times. Array Method (AM) is a scheduler to execute the task schedule in a set of clusters; this method was programmed sequentially, analyzed and tested using real and synthetic application workloads in previous work. Analyzing the proposed designs of this method in this research work, the parallelization of the method is extended using hybrid OpenMP and MPI programming in a server farm and using a set of geographically distributed clusters; at the same time, a novel method for searching free resources in clusters using Lévy random walks is proposed. Synthetic and real workloads have been experimented with to evaluate the performance of the new parallel schedule and compare it to the sequential schedule. The metrics of makespan, waiting time, quality of assignments and search for free resources were evaluated; the results obtained and described in the experiments section show a better performance with the new version of the parallel algorithm compared to the sequential version. By using the parallel approach with hybrid programming applied to the extraction of characteristics of the PTGs, applied to the search for geographically distributed resources with Lévy random walks and applied to the metaheuristic used, the results of the metrics are improved. The makespan is decreased even when the loads increase, the times of the tasks in the waiting queue are decreased, the quality of assignments in the clusters is improved by causing the tasks with their subtasks to be assigned in the same clusters or in cluster neighbors and, finally, the searches for free resources are executed in different geographically distributed clusters, not sequentially.

Visual Thinking Strategies—Theory and Applied Areas of Insertion

Twenty-first century learners live in a highly visual world, being constantly surrounded by visual, technologized stimuli and the educational system, be it general or specialized—as in the case of higher education—needs to creatively meet and answer these learning requirements. Among the plethora of new or updated approaches, the concept of Visual Thinking Strategies (VTS) designed by Philip Yenawine and Abigail Housen has gained extensive popularity due to its capacity of being adapted to a wide range of specialized fields with notable improvement results. The main purpose of the present review is to synthetically and critically present relevant scientific work related to the application of the VTS procedure and to further identify possible study areas that would highly benefit from the insertion of this procedure. Thus, the theoretical perspective is tackled from a cognitive-psychology standpoint, followed by a rendering of the research variety in applied VTS contexts pertaining to different study domains, as well as online VTS web clustering. These contexts have shown a predilection for VTS usage, which improves higher-education students’ specialized vocabulary and speaking skills in the same target language, but not in the case of ESP (English for specific purposes) classes.

Multi-Layer Web Services Discovery Using Word Embedding and Clustering Techniques

We propose a multi-layer data mining architecture for web services discovery using word embedding and clustering techniques to improve the web service discovery process. The proposed architecture consists of five layers: web services description and data preprocessing; word embedding and representation; syntactic similarity; semantic similarity; and clustering. In the first layer, we identify the steps to parse and preprocess the web services documents. In the second layer, Bag of Words with Term Frequency–Inverse Document Frequency and three word-embedding models are employed for web services representation. In the third layer, four distance measures, namely, Cosine, Euclidean, Minkowski, and Word Mover, are considered to find the similarities between Web services documents. In layer four, WordNet and Normalized Google Distance are employed to represent and find the similarity between web services documents. Finally, in the fifth layer, three clustering algorithms, namely, affinity propagation, K-means, and hierarchical agglomerative clustering, are investigated for clustering of web services based on observed similarities in documents. We demonstrate how each component of the five layers is employed in web services clustering using randomly selected web services documents. We conduct experimental analysis to cluster web services using a collected dataset consisting of web services documents and evaluate their clustering performances. Using a ground truth for evaluation purposes, we observe that clusters built based on the word embedding models performed better than those built using the Bag of Words with Term Frequency–Inverse Document Frequency model. Among the three word embedding models, the pre-trained Word2Vec’s skip-gram model reported higher performance in clustering web services. Among the three semantic similarity measures, path-based WordNet similarity reported higher clustering performance. By considering the different word representations models and syntactic and semantic similarity measures, we found that the affinity propagation clustering technique performed better in discovering similarities among Web services.

FPGA implementation of RDMA for ATLAS readout with FELIX at high luminosity LHC

The FELIX system is used as an interface between front-end electronics and commodity hardware in the server farm. FELIX is using RDMA through RoCE to transmit data from its host servers to the software readout driver using off-the-shelf networking equipment. RDMA communication is implemented using software on both end of the links. Exploring opportunities to improve data throughput as part of the high luminosity LHC upgrade, an implementation for RDMA support in the front-end FELIX FPGA is being developed. We present a proof-of-concept RDMA FPGA implementation, which will help inform the design of the FELIX platform for high luminosity LHC.

Edge Intelligence Empowered Dynamic Offloading and Resource Management of MEC for Smart City Internet of Things

Internet of Things (IoT) has emerged as an enabling platform for smart cities. In this paper, the IoT devices’ offloading decisions, CPU frequencies and transmit powers joint optimization problem is investigated for a multi-mobile edge computing (MEC) server and multi-IoT device cellular network. An optimization problem is formulated to minimize the weighted sum of the computing pressure on the primary MEC server (PMS), the sum of energy consumption of the network, and the task dropping cost. The formulated problem is a mixed integer nonlinear program (MINLP) problem, which is difficult to solve since it contains strongly coupled constraints and discrete integer variables. Taking the dynamic of the environment into account, a deep reinforcement learning (DRL)-based optimization algorithm is developed to solve the nonconvex problem. The simulation results demonstrate the correctness and the effectiveness of the proposed algorithm.

Delay-Join the Shortest Queue Routing for a Parallel Queueing System with Removable Servers

We introduce a new class of policies called “delay-join the shortest queue (delay-JSQ)” for use in parallel processing networks with removable servers. When jobs arrive to the system while all servers are on, jobs should be routed to the shortest queue. However, when servers are off, they take a random time to turn back on, which we allow to occur only when the number of jobs in each of the nonempty queues exceeds a fixed threshold. This new class of policies balances the load among all servers that are currently on and balances the capacity by keeping servers off until they are needed. A detailed numerical study shows that at moderate loads (where server farms and increasingly manufacturing facilities operate), delay-JSQ outperforms JSQ by up to 80%. In addition, it does so without precise knowledge of the input parameters and even when the input process is nonstationary.

Load Balancing Based on Closed Loop Control Theory (LBBCLCT): A Software Defined Networking (SDN) powered server load balancing system based on closed loop control theory

AbstractThis article comprises work concerned with proposing a scheme for intercepting network traffic and directs that traffic to servers in Software defined network model. Several studies have explored the effects of average response time of user requests with different load balancing algorithms. Server load balancing is a well‐recognized issue which necessitates a better approach by doing deeper understanding of the network parameters. The implications of the mean response time of client requests are investigated in this article by having a closer look on changing the important network parameter namely probing time (time to investigate the server farm in order to gain access to server data) using the concept of closed loop control theory. Our LBBCLCT proves to be effectively distributes the user requests to the most appropriate server by timely changing the server even when the traffic exhibits inconsistency. The results reveal that LBBCLCT outperforms other methods such as round robin, LBBSRT, and SD‐WLB in terms of processing the client requests with a reduced mean server response time. LBBCLCT showed the improvement of average response time by 46.2%, 25.6%, and 43.59% over round robin, LBBSRT, and SDWLB respectively.

Secure authentication and session key management scheme for Internet of Vehicles

AbstractInternet of Vehicles (IoV) is an internet enabled connected network of vehicles that provide services more effectively to its users via data sharing and communication with other vehicles and objects in the network. Use of fog servers to create a vehicular cloud which is connected to a primary cloud server is another effective technique that gives better performance by achieving better cooperation between the vehicular cloud and the internet. However, due to dependence on wireless technologies for communication, there is a need for a secure IoV network that can authenticate each device connecting to the network and can further secure the communication among these devices. This article introduced a secure authentication and session key management scheme for IoV (SASM‐IoV) working in fog computing‐based IoV network. SASM‐IoV first authenticates objects before entering into the network and then creates session keys for further communication to be secure. Devices registered to the network are assigned temporary identities, further used in the authentication process, thereby making our scheme more secure. SASM‐IoV protocol is effective against many security attacks such as impersonation, man‐in‐the‐middle, replay attack, and many more. Informal security analysis performed over SASM‐IoV shows that it is resilient against most security attacks. We have also performed a formal security analysis of the scheme using the widely accepted AVISPA tool. Network simulation and communication cost comparisons show that SASM‐IoV is better in terms of different network parameters and provides better security with lower costs when compared to similar security schemes.

Public Key Encryption with Authorized Equality Test on Outsourced Ciphertexts for Cloud-Assisted IoT in Dual Server Model

In cloud computing, the outsourced data face many privacy and security threats. To allow the cloud server to perform comparison, search, and classification on outsourced ciphertexts while simultaneously providing privacy guarantee, the encryption method that supports the ciphertext equality test is considered as a promising way. Users are able to authorize the cloud server to conduct the ciphertext equality test, so that two ciphertexts can be determined whether they encrypt the same message without being decrypted. In this process, users do not need to retrieve, decrypt, and then perform comparison on data; thus, the computing and communication efficiency can be greatly improved, and the privacy of user data can be guaranteed at the cloud server side. However, existing encryption schemes supporting authorized ciphertext equality test in the single server model cannot resist the keyword guessing attacks, and the solutions in the dual server model do not provide simultaneous authorization on two servers. To address these issues, this paper proposes a public key encryption scheme supporting authorized equality test on ciphertexts in the dual server model (PKE-AUT), where the primary server and secondary server must get the authorization from users before performing a sequential equality test on ciphertexts. Security and performance analysis demonstrate that the proposed PKE-AUT scheme not only guarantees the privacy of user data and authorization but also is practical in cloud-assisted IoT-related applications.

Insight: futurology

Abstract From chicken farms to server farms: Matt Lucas FBCS looks at what the poultry industry can teach us about effcient computing.

Optimal pricing in multi server systems

We study optimal service pricing in server farms where customers arrive according to a renewal process and have independent and identical (i.i.d.) exponential service times and i.i.d. valuations of the service. The service provider charges a time varying service fee aiming at maximizing its revenue rate. The customers that find free servers and service fees lesser than their valuation join for the service else they leave without waiting. We consider both finite server and infinite server farms. We solve the optimal pricing problems using the framework of Markov decision problems. We show that the optimal prices depend on the number of free servers. We propose algorithms to compute the optimal prices. We also establish several properties of the optimal prices and the corresponding revenue rates in the case of Poisson customer arrivals. We illustrate all our findings via numerical evaluation.

Food for thought: Commentary on Burnette et al. (2021) "Concerns and recommendations for using Amazon MTurk for eating disorder research".

Burnette et al. aimed to validate two eating disorder symptom measures among transgender adults recruited from Mechanical Turk (MTurk). After identifying several data quality issues, Burnette et al. abandoned this aim and instead documented the issues they faced (e.g., demographic misrepresentation, repeat submissions, inconsistent responses across similar questions, failed attention checks). Consequently, Burnette et al. raised concerns about the use of MTurk for psychological research, particularly in an eating disorder context. However, we believe these claims are overstated because they arise from a single study not designed to test MTurk data quality. Further, despite claiming to go "above and beyond" current recommendations, Burnette et al. missed key screening procedures. In particular, they missed procedures known to prevent participants who use commercial data centers (i.e., server farms) to hide their true IP address and complete multiple surveys for financial gain. In this commentary, we outline key screening procedures that allow researchers to obtain quality MTurk data. We also highlight the importance of balancing efforts to increase data quality with efforts to maintain sample diversity. With appropriate screening procedures, which should be preregistered, MTurk remains a viable participant source that requires further validation in an eating disorder context.

Adaptive Server Load Balancing in SDN Using PID Neural Network Controller

Web service applications are increasing tremendously in support of high-level businesses. There must be a need of better server load balancing mechanism for improving the performance of web services in business. Though many load balancing methods exist, there is still a need for sophisticated load balancing mechanism for not letting the clients to get frustrated. In this work, the server with minimum response time and the server having less traffic volume were selected for the aimed server to process the forthcoming requests. The Servers are probed with adaptive control of time with two thresholds L and U to indicate the status of server load in terms of response time difference as low, medium and high load by the load balancing application. Fetching the real time responses of entire servers in the server farm is a key component of this intelligent Load balancing system. Many Load Balancing schemes are based on the graded thresholds, because the exact information about the network flux is difficult to obtain. Using two thresholds L and U, it is possible to indicate the load on particular server as low, medium or high depending on the Maximum response time difference of the servers present in the server farm which is below L, between L and U or above U respectively. However, the existing works of load balancing in the server farm incorporate fixed time to measure real time response time, which in general are not optimal for all traffic conditions. Therefore, an algorithm based on Proportional Integration and Derivative neural network controller was designed with two thresholds for tuning the timing to probe the server for near optimal performance. The emulation results has shown a significant gain in the performance by tuning the threshold time. In addition to that, tuning algorithm is implemented in conjunction with Load Balancing scheme which does not tune the fixed time slots.

Optimized load balancing using software-defined networking (SDN)

In a cloud environment, handling user service requests and providing the requested resources fairly is critical. Load balancing is important to distribute service requests fairly to an unloaded server and to dynamically maintain load balancing across server farms. In conventional internet protocol (IP) networks, maintaining load balancing is a stubborn and non-adaptable task due to the forwarders' lack of global topology representation. Software-Defined Networking (SDN) provides centralized decision-making for any topological changes to manage changes dynamically. To solve the above problem, we propose a new server-side load balancing strategy that provides an efficient and effective server management scheme for SDN Open Flow networks. Experiments were done on the Ryu controller and the Mininet emulator showed that the performance was better than what was already available.

Towards A Strategic Approach Of Covid-19 Cluster Web Mapping In Malaysia

The world was shocked by an unprecedented outbreak caused by coronavirus disease 2019 (COVID-19). In Malaysia, it started with the largest number of COVID-19 cases with the first wave of infection on 25 January 2020. The objectives of this paper are to obtain the perspective of the respondents about the need for web-mapping in the form of mapping the geospatial data in Malaysia and to visualize the current online datasets of COVID-19 disease case clusters. The study area would cover the entire Malaysia since a rapidly increasing number of citizens were affected by this virus. To be specific, this study focused on the active clusters of COVID-19 in Malaysia. The data were freely shared in real-time by referring to the Ministry of Health (MOH) channel. The hotspots map were explored using the Map Editor by Cloud GIS. The approach has been illustrated using a dataset of whole Malaysia which are locally transmitted confirmed cases in four phases of COVID-19 wave in Malaysia. This study is significant to raise public awareness of the virus, especially among Malaysian citizens. It can provide an accurate estimation of the cluster tracking of the COVID-19 system by using geospatial technology. Therefore, people are more concerned and followed all the Standard Operating Procedure (SOP) provided by the government to prevent the spread of COVID-19.